Beamforming System Based on Delay Distribution Model Using High Frequency Phase Difference

1. An acoustic signal processing system, comprising: an input module configured to: receive at least two acoustic signals via at least two acoustic sensors; convert the at least two acoustic signals into at least two channels of analog signals and, subsequently, at least two channels of digital signals; a phase-difference module configured to: convert the at least two channels of digital signals to at least two channels of frequency transforms; and calculate phase differences between two selected channels; wherein each frequency transform comprises a plurality of complex numbers; wherein each complex number corresponds to a frequency bin; and wherein each phase difference is defined within (−π, π) or (−180 degrees, 180 degrees); a delay distribution module configured to: for each frequency bin, derive ambiguous delays from the phase differences by adding or subtracting multiples of 2π or 360 degrees; keep the ambiguous delays within a valid delay range as candidate delays; for each candidate delay, add a spread function centering each candidate delay to form a delay distribution function; wherein the valid delay range is predetermined according to a maximum acoustic propagation time delay between the at least two acoustic sensors plus a headroom; a delay estimation module configured to make a final delay estimation based on the delay distribution function; and a delay-and-sum module configured to align one of the two selected channels according to the final delay estimation to obtain a signal of interest.

2. The system of claim 1 , wherein the delay distribution module is further configured to apply a frequency-dependent weighting function to the spread function.

3. The system of claim 1 , wherein the delay distribution module is configured to apply an energy-dependent temporal adapting scheme to smooth the delay distribution function before making the final delay estimation.

4. The system of claim 1 , wherein the phase-difference module is configured to use Fourier transform to convert the at least two channels of digital signals to at least two channels of frequency transforms.

5. The system of claim 1 , further comprising a digital signal processor to implement the phase-difference module, the delay distribution module, the delay estimation module, and the delay-and-sum module.

6. The system of claim 1 , further comprising software codes executed in a general purpose processor to implement the phase-difference module, the delay distribution module, the delay estimation module, and the delay-and-sum module.

7. An acoustic signal processing system, comprising: a microphone interface circuit configured for coupling to first and second acoustic sensors to receive first and second acoustic signals from a same acoustic signal source and to convert the first and second acoustic signals to first and second analog signals, respectively; an analog-to-digital converter configured to receive the first and second analog signals and to generate first and second digital signals, respectively; and a signal processing circuit configured to receive the first and second digital signals and to determine a delay between the first and second digital signals, wherein the signal processing circuit comprises: a phase-difference module configured to: transform the first and second digital signals to provide first and second frequency domain signals; calculate a first set of phase differences between the first and second frequency domain signals at a plurality of selected frequencies; wherein each phase difference is a value defined within (−π, π) or (−180 degrees, 180 degrees); a delay distribution module configured to: derive a second set of phase differences by adding and subtracting multiples of 2π or 360 degrees to each of the first set of phase differences; derive a plurality of candidate delay times at the plurality of selected frequencies based on the first set of phase differences and the second set of phase differences; for each candidate delay, add a spread function centering each candidate delay to form a delay distribution function; and determine a histogram of the plurality of candidate delay times; a delay estimation module configured to determine an estimated delay time by selecting a delay time having a maximum count in the histogram of the plurality of candidate delay times; and a delay-and-sum module configured to align the first and second digital signals according to the estimated delay time to obtain a signal of interest.

8. The system of claim 7 , wherein the delay distribution module is further configured to apply a frequency-dependent weighting function to the spread function.

9. The system of claim 8 , wherein the delay distribution module is configured to apply an energy-dependent temporal adapting scheme to smooth the delay distribution before determining the estimated delay time.

10. The system of claim 7 , wherein the phase-difference module is configured to use Fourier transform to transform the first and second digital signals to provide first and second frequency domain signals.

11. The system of claim 7 , wherein the delay distribution module is further configured to, for each frequency bin, select the plurality of candidate delay times within a valid delay range as candidate delays; wherein valid delay range is predetermined according to the maximum acoustic propagation time delay between the first and second acoustic sensors plus a headroom.

12. The system of claim 7 , wherein the signal processing circuit comprises a digital signal processor.

13. The system of claim 7 , wherein the first and second acoustic sensors are disposed apart by a spacing greater than a half wavelength of the acoustic signals.

14. A method, comprising: receiving acoustic signals from an acoustic signal source using first and second acoustic sensors, the first acoustic sensor receiving a first acoustic signal and the second acoustic sensor receiving a second acoustic signal; converting the first and second acoustic signals into first and second digital signals; transforming the first and second digital signals into first and second frequency domain signals; determining a first set of phase differences between the first and second frequency domain signals, the first set of phase differences being defined within (−π, π) or (−180 degrees, 180 degrees); determining a second set of phase differences by adding and subtracting multiples of 2π or 360 degrees to each of the first set of phase differences; determining a plurality of candidate delay times between the first and second frequency domain signals at a plurality of frequencies based on the first set of phase differences and the second set of phase differences; for each candidate delay, adding a spread function centering each candidate delay to form a delay distribution function; and selecting an estimated delay time from the plurality of candidate delay times, the estimated delay time being associated with a largest number of the plurality of frequencies, wherein the estimated delay time is associated with an estimated distance between the first and second acoustic sensors.

15. The method of claim 14 , further comprising applying a frequency-dependent weighting function to the spread function.

16. The method of claim 14 , further comprising apply an energy-dependent temporal adapting scheme to smooth the delay distribution function before making the final delay estimation.

17. The method of claim 14 , further comprising using a Fourier transform to transform the first and second digital signals into first and second frequency domain signals.

18. The method of claim 14 , wherein the first and second acoustic sensors are disposed apart by a spacing greater than a half wavelength of the acoustic signals.

19. The method of claim 14 , further comprising: determining a histogram of the plurality of candidate delay times; and determining an estimated delay time by selecting a delay time having a maximum count in the histogram of the plurality of candidate delay times.